By now, I think we all have probably figured out that Ken Pomeroy, the brains behind KenPom.com, is the best numbers wizard currently analyzing college hoops. If you haven’t figured out that you should be mining his site for score projections that are significantly different from Vegas betting lines, then you are missing out on an edge when it comes to winning money.
I bring this all up because of a study that Ken did last week, one that I think is quite relevant as we get ready to ramp up for another season of betting more than we can afford on college hoops. Essentially what happened is that Ken went through and tried to determine what box score stat had the biggest impact on winning, and he did this by looking at how win probabilities change for conference games that were tied at the half.
The answer is first half three-point percentage. The team that shot worse from three received a 7 percent bump in win probability relative to their opponent if they shot a lower percentage from beyond the arc in the first 20 minutes.
On the surface, this might seem a bit strange, but logically it makes perfect sense. I don’t have any data to actually back this up, but outside of a star player getting into foul trouble and having to sit for a significant amount of time, three-point shooting is the one aspect of basketball where it is the easiest to have an outlier performance, and where said outlier performance has the biggest impact on the result of a game, or a half.
Think about it like this. Let’s say that your team shoots 40 percent from three on the season and averages 10 threes per half of basketball. That means that in a given 20 minute period, we can expect your team to shoot 4-for-10 from beyond the arc. If you shoot 6-for-10 from three in the first half of a game, that is an extra six points that we cannot project for the second half. The opposite is true is you shoot 2-for-10 from three in the first half.
Now imagine that the team that shot 6-for-10 from three is playing the team that shot 2-for-10 from three and that the score is tied at the half. That’s a 12-point swing on four actions in a game that probably had around 70-75 possessions in the first half. Assuming that three-point shooting will regress to the mean – meaning that both teams return to their average of 4-for-10 shooting – the team that shot worse from three in the first half should be expected to have a better second half performance.
Logically, that all makes sense.
But it is nice for Ken to prove it to us.
And while context will be important in every situation where we use this as a way to try and make money next season, do remember that the team with a lower three-point shooting percentage was the one stat that impacted second half win probabilities the most.
So keep that in mind next season when you are live-betting games at the half.